PrimeGrid Races 2018

[Ken g6]

Sorry, no stats today. My main internet is down and my phone doesn't have the necessary scripts.

 

[StefanR5R]

Internet down while two races are going on?

Day 3 stats, from the UTC 12:00 mark:

Rank___Credits____Username
2______3739086____xii5ku
48_____344981_____crashtech
50_____314163_____Howdy2u2
101____132718_____zzuupp
125____106312_____Ken_g6
182____53163______bbhaag
160____71929______Orange Kid
260____6680_______biodoc

Rank__Credits____Team
2_____11010393___Czech National Team
3_____10956048___SETI.Germany
4_____8752081____Sicituradastra.
5_____4769032____TeAm AnandTech
6_____2885507____BOINC@Poland
7_____2441937____AMD Users
8_____2134137____Crunching@EVGA

Graph from UTC 19:30:

(source)

 

[Ken g6]

For now I'm calling it forced bunkering. If the work runs out I would have to call it something else.

 

[Ken g6]

I guess I should provide some stats this morning, now that my Internet is back up. Thanks to @StefanR5R for doing the stats yesterday.

Day 4 stats:

Rank___Credits____Username
2______5298375____xii5ku
53_____449296_____Howdy2u2
69_____344980_____crashtech
112____179836_____zzuupp
157____112998_____Ken_g6
161____106741_____bbhaag
175____91715______Orange Kid
306____6680_______biodoc

Rank__Credits____Team
2_____16104194___Czech National Team
3_____15706158___SETI.Germany
4_____13054551___Sicituradastra.
5_____6590626____TeAm AnandTech
6_____4180732____BOINC@Poland
7_____3556170____AMD Users
8_____2888754____Crunching@EVGA

We're not exactly going at full speed, but I don't think anyone else is trying hard to catch us.

 

[Orange Kid]

Thanks for the stats update.

 

[TennesseeTony]

Indeed, thanks for the stats.

And, I just have to give a shout out to Stefan, Howdy, and crashtech for their impressive scoring! Two challenges at once sucks, so thanks (to all participating) for keeping the TeAm in good standing here in PG, and respected globally!

 

[Howdy]

Appreciative of the shout, but the Star of the show is Stefan and his scoring!! Crashtech would be 20-25 spots above me if he (I'm assuming) didn't divert his attention elsewhere. Kudos to him for the redirect!! I myself stayed on/ went back to PG since my cr_app is pitiful on the Sprint.
All of us together though make a heck of a TeAm!!

 

[StefanR5R]

Thanks guys for tolerating my fooling around at PrimeGrid till the end, even though our team ranking here was already safer than at the FB sprint. (It paid off for my individual ranking. Good for the TeAm that my affiliation is shown in the full table at the PrimeGrid site. Sadly the team affiliation won't be stated in the PG forum post which announces winners when validations are complete.)

 

[TennesseeTony]

Well done on hanging onto 2nd!!! Scott was creeping up on you it looked like and had me worried on your behalf.

 

[Ken g6]

Preliminary final stats:

Rank___Credits____Username
2______6410969____xii5ku
57_____523804_____Howdy2u2
87_____338344_____crashtech
111____227184_____zzuupp
146____160785_____bbhaag
168____126460_____Ken_g6
193____98480______Orange Kid
324____6680_______biodoc

Rank__Credits____Team
2_____20791542___Czech National Team
3_____19167185___SETI.Germany
4_____15820375___Sicituradastra.
5_____7892710____TeAm AnandTech
6_____5052279____BOINC@Poland
7_____4399894____AMD Users
8_____3585160____Crunching@EVGA

Wow, another great finish thanks to @StefanR5R! Too bad that sprint was so distracting, but at least I got a few more WUs in with my laptop.

 

[Howdy]

Umm you do see/realize that Stefan with his score had 5th wrapped up by himself !!

 

[crashtech]

Well, we might at least be encouragement for him?

 

[ao_ika_red]

*ICYMI*
PrimeGrid: New LLR apps

We have upgraded the version of LLR used on all of our LLR projects to LLR v.3.8.21.

As with the previous version of LLR, it is available in 32 and 64 bit versions for Windows and Linux, and 64 bits for Mac OS.

This release has two major features:

* Support for AVX/FMA3 on AMD Ryzen CPUs. You can expect about a 10% increase in speed on Ryzen processors as compared to earlier versions of LLR.

* A rare bug that sometimes caused errors when using multithreading has been corrected.

If you are using app_info.xml (aka anonymous platform) please upgrade to the latest LLR. The wrapper is unchanged. Upgrading is not mandatory, but is strongly recommended.

If you are not using app_info.xml, you will automatically get the new version with your next LLR task and no action is necessary on your part. If you're not sure if you're using app_info.xml you're almost certainly not using it, and no action is necessary.

 

[Ken g6]

Also please note that using an app_config.xml is not the same as using app_info.xml. Your apps will automatically update if you are just using app_config.

 

[Howdy]

I fit in with this line:

If you are not using app_info.xml, you will automatically get the new version with your next LLR task and no action is necessary on your part. If you're not sure if you're using app_info.xml you're almost certainly not using it, and no action is necessary.

 

[StefanR5R]

On April 9, results of the Year of the Dog challenge in 321-LLR, March 16-23, became final.

Deductions after March 23 due to invalid results:
1 ... Aggie The Pew ...................... -4,977 ....... -1 WU ....... -0.01 %
2 ... Czech National Team ......... -93,951 ..... -19 WUs ..... -0.29 %
3 ... SETI.Germany ....................... -9,856 ....... -2 WUs ..... -0.03 %
4 ... Sicituradastra. .................... -39,694 ....... -8 WUs ..... -0.17 %
5 ... L'Alliance Francophone ..... -19,805 ....... -4 WUs ..... -0.13 %
6 ... TeAm AnandTech .............. -4,941 ....... -1 WU ....... -0.04 %
7 ... Team China ........................ -49,199 ..... -10 WUs ..... -0.46 %
8 ... Crunching@EVGA ................ -4,971 ....... -1 WU ....... -0.06 %

By the metric of the quality of returned results, we came in as a close 3rd. :-)


On April 26, results of the Mathematics Awareness Month challenge in ESP-LLR, April 3-8, became final.

Deductions after April 8 due to invalid results:
1 ... Aggie The Pew ...................... -6,726 ....... -1 WU ....... -0.03 %
2 ... Czech National Team ......... -66,654 ..... -10 WUs ..... -0.32 %
3 ... SETI.Germany .............................. -0 ....... -0 WUs ..... -0.00 %
4 ... Sicituradastra. .................... -33,249 ....... -5 WUs ..... -0.21 %
5 ... TeAm AnandTech ............ -13,247 ....... -2 WUs ..... -0.17 %
5 ... BOINC@Poland ........................... -0 ....... -0 WUs ..... -0.00 %
7 ... AMD Users .................................. -0 ....... -0 WUs ..... -0.00 %
8 ... Crunching@EVGA ................ -6,621 ....... -1 WU ....... -0.19 %

Only 5th by quality this time, need to get back to earlier form in the next challenge. ;-)


Note, I didn't count WUs which were already invalidated before the end of the challenge, because these went mostly undetected in the stats available to me.

 

[TennesseeTony]

Back to the top....

There is a PG event coming up this week, "To participate in the Challenge, please select only the Sierpinski / Riesel Base 5 LLR (SR5) project in your PrimeGrid preferences section. The challenge will begin 14th June 2018 00:00 UTC and end 19th June 2018 00:00 UTC."

 

[Ken g6]

I've been testing PrimeGrid tasks, since this will be my first time running a 14-core processor there.

It looks like one WU at a time would run in about 48 minutes or so. But if I break it into two, they run in about 73 minutes, or 38 minutes each. Maybe a good rule of thumb is, if the WU would complete in less than an hour, add another WU? Or do you think I should try 3 WUs at a time?

 

[StefanR5R]

My guess is that Haswell-EP and Broadwell-EP scale similarly. I've got 14 cores too, though at slightly higher clocks and in a dual-socket board.

I'm a bit late (though with recent average of 10 CPU hours not too late), but I just now started to dig out my results of the tests that I did before the llrESP race, in order to set up some tests with a fixed WU. Tests with a little 4C/8T host are already underway, though these will probably lack surprises.

 

[StefanR5R]

Uh oh:

• I haven't run PrimeGrid LLR since they updated the application. Therefore I loaded one task on most of the computers which I plan to use in the next challenge, just in order to get the current wrapper and LLR exe. While this took only minutes (as expected) on all but one computer, one is taking 1.5 hours to download the LLR exe. Strange.

• The first llrSR5 task which I tested took 2,910 seconds. But in my task list at the primegrid website, it is listed with 3,124 seconds. And that's even though I downloaded only this single task. Conclusion: The task list at primegrid.com is completely useless for performance tracking of multithreaded applications.
  Edit,
  same with three other hosts. Each downloaded and ran only a single task: The run time shown in the task list at the web site was always wrong (too long).

 

[crashtech]

I'm assuming that there are slightly different "sweet spots" for differing architectures and platforms. Of particular interest to many of us will be the proper settings for Ryzen and Threadripper. I wish I had the time, knowledge, and organizational skills to do some of my own testing.

 

[TennesseeTony]

And.....it's go time.

 

[crashtech]

I'm in!

 

[Markfw]

Back to the top....

There is a PG event coming up this week, "To participate in the Challenge, please select only the Sierpinski / Riesel Base 5 LLR (SR5) project in your PrimeGrid preferences section. The challenge will begin 14th June 2018 00:00 UTC and end 19th June 2018 00:00 UTC."

OK, I am in. But one day per task ?????

16 tasks at a time on a 2700x@4.2 ghz, memory 3600 with subtimings set, my fastest single-core box, but with 8c/16t

 

[zzuupp]

ah! poop!

I knew I wouldn't be around near GO! time. Starting at lunch was less than successful. I was hoping on good timing so that round2 would fully count. Alas, very little luck.

I'm in, but I'm very late.